#linuxcnc Logs

Apr 11 2023

12:30 AM -!- #linuxcnc mode set to +v by ChanServ
12:56 AM Deejay: moin
01:14 AM Vvega[m]: Haha
01:16 AM Vvega[m]: Right well , got it all working , I might need to put a ferrite on the voltage input … or used a screened wire , cause I’m getting some groaning that shuts up when you disconnect the an connector
01:16 AM Vvega[m]: Just need to work out the spindle scaling and sync
04:28 AM -!- #linuxcnc mode set to +v by ChanServ
04:45 AM Tom_L: morning
04:52 AM -!- #linuxcnc mode set to +v by ChanServ
04:52 AM JT-Cave: morning
04:53 AM Tom_L: i tried mesact2 on armhf a 2nd time and still got that error
05:01 AM JT-Cave: odd I have it running on armhf now with no error
05:08 AM JT-Cave: are you running from a terminal?
06:49 AM Tom_L: yes
06:49 AM Tom_L: python3 mesact
07:25 AM JT-Cave: Tom_L, run python3 in a terminal then
07:25 AM JT-Cave: from PyQt5 import QtCore
07:25 AM JT-Cave: QtCore.qVersion()
07:25 AM JT-Cave: my armhf is 5.15.2, this pc is 5.15.8
07:28 AM beef[m]: just your friendly annual reminder to ensure your postprocessor is actually generating G2/G3 arcs by the way
07:29 AM beef[m]: i never paid much attention to the gcode since it worked, until resolving a tangentially related problem and he mentioned "hey, where dafuq are your G2/G3 arcs??"
08:34 AM skunkworks[m]: beef (@_discord_617572522933616647:jauriarts.org) you might want to run some speed comparisons. (assuming you are doing any 'high speed' machining. It may be slightly faster as short line segments.
08:35 AM beef[m]: that would be all based on accels though, no?
08:35 AM ZincBoy[CAON][m]: I don't bother with arc fitting for complex tool paths. If it is a simple 2d path then fine. But HSM I just use line segments.
08:59 AM skunkworks[m]: beef eh.. arc to arc blending might not be as fast.. But - it has been a while and I might be miss-remembering
09:34 AM JT-Cave: Tom_L, seems that issue has something to do with the python and or pyqt5 version
09:34 AM JT-Cave: https://stackoverflow.com/questions/70795081/python-pyqt5-attributeerror-qtabwidget-object-has-no-attribute-settabvisibl
09:55 AM JT-Cave: void QTabWidget::setTabVisible(int index, bool visible)
09:55 AM JT-Cave: If visible is true, the page at position index is visible; otherwise the page at position index is hidden. The page's tab is redrawn appropriately. This function was introduced in Qt 5.15.
01:03 PM JT-Woodshop: the jig has a slight flaw
01:04 PM Tom_L: i'm afraid if i update buster on the rpi, it will break the usb so i'm still running the original install on it
01:04 PM -!- #linuxcnc mode set to +v by ChanServ
01:05 PM -!- #linuxcnc mode set to +v by ChanServ
01:06 PM Tom_dev: >>> QtCore.qVersion()
01:06 PM Tom_dev: '5.11.3'
01:06 PM Tom_dev: >>>
01:08 PM JT-Woodshop: try updating python3-pyqt5
01:09 PM JT-Woodshop: nap time
01:09 PM Tom_dev: how can i update just that one package?
01:09 PM Tom_dev: E: The update command takes no arguments
01:10 PM JT-Shop: sudo apt install python3-pyqt5
01:10 PM JT-Shop: that should update it
01:10 PM Tom_L: nope
01:10 PM Tom_dev: 0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.
01:11 PM JT-Shop: python3-pyqt5 is already the newest version (5.15.9+dfsg-1).
01:11 PM JT-Shop: what does yours say
01:11 PM Tom_dev: python3-pyqt5 is already the newest version (5.11.3+dfsg-1+rpi1+b3).
01:12 PM Tom_L: you're running an rpi os
01:12 PM Tom_L: this is buster
01:12 PM JT-Shop: this is x86 but I have a rpi down in the cave
01:12 PM JT-Shop: 1417 packages can be upgraded. Run 'apt list --upgradable' to see them.
01:13 PM JT-Shop: but I'm afraid to upgrade this pc after the fiasco on the dev pc
01:13 PM JT-Shop: anyhow nap time
01:13 PM Tom_L: i am too because when i did on the rpi the usb quit working
01:13 PM Tom_L: lunch then a nap
01:28 PM seb_kuzm1nsky is now known as seb_kuzminsky
01:29 PM mrec: does anyone have experience with backlash? (I'm looking at the mechanical aspect).
01:30 PM mrec: I'm using the JMC IHSV57 motors here, there's some backlash depending on which direction I come from
01:30 PM mrec: the JMC statistics show up the same position regardless if I come from the left or right side
01:31 PM roycroft: everybody here has experienced backlash in one form or another :)
01:31 PM mrec: it's no real mechanical backlash but some kind of motor backlash, the servo pid will regulate back if I try to move the head
01:32 PM Unterhausen: are they steppers?
01:32 PM mrec: bldcs, closed loop
01:34 PM Unterhausen: not really backlash then. Possibly fixed by more tuning
01:34 PM Unterhausen: increase ff gains maybe?
01:36 PM mrec: the motor data says it's at the correct position
01:36 PM mrec: but it's weird.. if I turn the motor manually it always regulates back to that certain position
01:36 PM Unterhausen: how do you decide you're having a problem then?
01:37 PM mrec: I see it with the attached camera
01:38 PM Unterhausen: the clearpath motors allow you to see the error
01:39 PM Unterhausen: if you are relying on one bit output to tell you that it's in position, there will always be an error band
01:39 PM mrec: that's the point the motors say there is no error
01:39 PM Unterhausen: is there a manual online
01:39 PM mrec: https://i.snipboard.io/VnIMPX.jpg
01:39 PM Unterhausen: they don't want to be constantly erroring out
01:39 PM mrec: https://i.snipboard.io/2ORZUF.jpg
01:40 PM mrec: I can move over hundred times, it will come back to that position but it depends if I come from the left or from the right side
01:40 PM mrec: I cannot move the camera manually it's fixed, the motor insists on that position
01:41 PM mrec: I couldn't find a manual for the motor online which addresses such an issue
01:41 PM bjorkintosh: GIGO. it must be getting data from somewhere.
01:42 PM mrec: there's an encoderwheel in the motor, maybe I should attach my own counter to it
01:56 PM mrec: https://i.snipboard.io/CAVrPt.jpg
01:56 PM mrec: Position feedback bit is always 2106.99996948242 regardless if I come from the left or from the right side
02:05 PM mrec: I guess that offset is around 0.1mm, possibly it matches the raster settings
02:05 PM mrec: 500 pulses for 20mm
02:09 PM pcw-home: That seems pretty coarse
02:10 PM mrec: certainly, but it shouldn't be that inaccurate, especially since the wheel has 1000ppr (4x - 4000ppr)
02:11 PM mrec: it's not a CNC, but a pick and place machine
02:11 PM mrec: 0.04mm is good enough
02:12 PM pcw-home: but it may have to do with you hysteresis issue
02:14 PM mrec: considering that the wheel has 4000ppr (4x mode) that would speak against it.
02:14 PM pcw-home: depends on how it handles the step pulses
02:15 PM pcw-home: increasing the steps/mm might be worth a try
02:15 PM mrec: I would need a new microcontroller then
02:15 PM pcw-home: (could also improve smoothness)
02:16 PM mrec: the motor has some issues here and there if you look at the details. for an average application it's good enough.
02:17 PM mrec: https://www.jmc-motor.com/product/901.html
03:03 PM * JT[m] uploaded an image: (528KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/NMEFWrXfMEGDeKekaBvYumTV/20230411_150141.jpg >
03:03 PM * JT[m] uploaded an image: (291KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/aFludwQIZVOeoSUqLOQEOfAd/20230411_150156.jpg >
03:04 PM travis_farmer[m]: hey, you finished my bowl? ;-)
03:04 PM XXCoder: take that colorful wood chips and make small bowl using that and epoxy ;)
03:04 PM XXCoder: oh travis_farmer[m] you getting married? ;)
03:05 PM travis_farmer[m]: i am? wow, news to me... ;-)
03:05 PM XXCoder: jb is making it for marrage gift, and you said its yours? ;)
03:05 PM XXCoder: anyway bowl looks awesome!
03:06 PM travis_farmer[m]: i also added the wink face emogi... ;-)
03:06 PM XXCoder: yeah it shows up on irc.
03:06 PM rmu: mrec AC servo adjsut software ??
03:19 PM JT-Woodshop: thanks
03:19 PM JT-Woodshop: compressing the chips with some epoxy might make something interesting and might not
03:20 PM Unterhausen: etsy thinks wood chips in epoxy is interesting
03:21 PM XXCoder: yeah. its silly but might be interesting.
03:23 PM * roycroft thinks epoxy is an adhesive, and sometimes a protective finish, but never a design element
03:24 PM * JT[m] uploaded an image: (236KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/xzjwjltwVVFUzgWrNEIaemPB/20230411_152313.jpg >
03:24 PM JT[m]: New circle cutting jig has a flaw
03:56 PM acer is now known as _unreal_
04:00 PM Rhine_Labs: Sup all
04:01 PM XXCoder: JT[m]: what flaw?
04:04 PM JT-Shop: needs to be longer so you can start the cut with the guide in the slot fully
04:05 PM XXCoder: ahh yeah
04:05 PM JT-Shop: it was not obvious until I cut the disk out
04:06 PM XXCoder: yeah reality is best test of any design and ideas :D
04:06 PM JT-Shop: fortunately it's only a bit of wood
04:09 PM XXCoder: yeah
04:20 PM -!- #linuxcnc mode set to +v by ChanServ
04:29 PM roycroft: i think it's time for a class action suit against festool for false advertising
04:29 PM roycroft: festoolusa claim to have an official mascot dog - "sparky" - who is featured in almost every festool live episode
04:29 PM roycroft: yet i just received a shipment from festool
04:30 PM roycroft: it's the spanner that was missing from the router i recently purchased
04:30 PM roycroft: it came in a television-size box
04:30 PM roycroft: this is proof that the company is actually operated by cats, and there's no way their mascot is a dog
04:31 PM XXCoder: cats sure own a lot of shipping services
04:32 PM roycroft: yes
04:33 PM roycroft: and the dirty truth that has not come out yet is that covid-19 is not a chinese invention
04:33 PM roycroft: it was created by cats
04:35 PM XXCoder: an bird virus to slow down birds for cats?
04:35 PM roycroft: what has covid-19 done?
04:36 PM roycroft: 1. it has kept the cats' servants at home insted of their going away for 8 hours/day to work
04:36 PM roycroft: 2. it has significantly increased the number of boxed deliveries to their houses
04:39 PM JT-Woodshop: so to make a 16" guide for the circle jig I need to either cut a slightly wider bit from purpleheart or find that bigger chunk of hickory
04:45 PM JT-Woodshop: hmm the hickory is 1 1/2" x 2" x 36"...
04:47 PM JT-Woodshop: I think Leo has influenced me to use purpleheart
04:49 PM * JT-Woodshop found a Bunong folk song he likes Ra Ngol Ra Hang
04:53 PM -!- #linuxcnc mode set to +v by ChanServ
04:54 PM -!- #linuxcnc mode set to +v by ChanServ
04:56 PM roycroft: jt-woodshop: which festool extractor do you have?
04:57 PM roycroft: that pic of your bandsaw fence shows it, and it looks like a midi to me
04:57 PM JT-Woodshop: CT MIDI 1
04:58 PM roycroft: ok
04:58 PM roycroft: so the same as mine
04:59 PM roycroft: i've been really happy with mine so far
04:59 PM roycroft: if i start using that new router a lot i might wish i had one with more capacity, like a ct 36
04:59 PM roycroft: but i only go through a few bags/year as it is
05:00 PM roycroft: i can buy a lot of bags for the cost difference between a midi and a ct 36
05:00 PM roycroft: even at festool prices for the bags
05:04 PM JT-Woodshop: I use the reusable bag
05:05 PM JT-Woodshop: https://freemusicarchive.org/music/Les_Cartes_Postales_Sonores/NYEL_CHE_-_Folk_songs_of_the_Bunong_-_Vol1/Les_Cartes_Postales_Sonores_-_NYEL_CHE_-_Folk_songs_of_the_Bunong_-_Vol1_-_14_Ra_Ngol_Ra_Hang/
05:07 PM evie_kurt001[m]: I'll help 10 people to earn $40k within 72 hours but you will pay me 10% of your profit when you receive it. Note only interested people should apply, drop a message let's get started by asking (HOW)
05:07 PM evie_kurt001[m]: https://t.me/Eviekurt_admin
05:09 PM JT-Woodshop: Tom_L, how do you ban people?
05:21 PM roycroft: /mode #linuxcnc +b *!user@host
05:22 PM JT-Woodshop: me needs a new mouse here the right button is doa
05:22 PM roycroft: however, that spammer is in the matrix, so i don't know if the ban will work on irc
05:23 PM JT-Woodshop: prob right
05:26 PM roycroft: /ignore works though :)
05:32 PM JT-Woodshop: :)
05:56 PM Tom_L: JT-Woodshop, who?
05:56 PM JT-Shop: the spammer evie
05:57 PM Tom_L: guess i missed it
05:58 PM evie_kurt001[m]: I'll help 10 people to earn $40k within 72 hours but you will pay me 10% of your profit when you receive it. Note only interested people should apply, drop a message let's get started by asking (HOW)
05:58 PM evie_kurt001[m]: https://t.me/Eviekurt_admino
05:59 PM JT-Shop: Tom_L, there you go
05:59 PM -!- #linuxcnc mode set to +o by ChanServ
05:59 PM JT-Shop: thanks
05:59 PM Tom_L: wasn't a ban
06:00 PM Tom_L: gettin a little rusty..
06:05 PM -!- #linuxcnc mode set to -o by ChanServ
07:06 PM ithe2guu2795[m]: How do I want to configure this servo drive running in pulls/dir mode as a spindle motor?
07:06 PM * ithe2guu2795[m] uploaded an image: (1121KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/zsOMEZhCSaOhKFcIljmzOerU/IMG_1379.jpg >
07:07 PM djdelorie: what are you interfacing the spindle to? I mean, on the PC side
07:10 PM ithe2guu2795[m]: I have it hooked up to a 7i96S
07:10 PM ithe2guu2795[m]: The encoder is wired up, and they want to control it using step/ DIR
07:11 PM ithe2guu2795[m]: s/they/I/
07:12 PM djdelorie: Do you have a specific reason for using step/dir? That's usually used for axis control, not spindle. THe 7i96S has an analog spindle out as well as a serial port that can be (maybe) used for modbus
07:13 PM djdelorie: I use modbus to run my spindle
07:18 PM ithe2guu2795[m]: I could not figure out how to set it up any other way
07:18 PM ithe2guu2795[m]: how do you think i should set it up?
07:19 PM ithe2guu2795[m]: like i could not find the pins to do it either of those ways
07:19 PM djdelorie: the easy way is to use the analog interface on the 7i96 - on the servo, select "speed servo" mode and Analog command input
07:19 PM ithe2guu2795[m]: ok
07:20 PM djdelorie: on the 7i96, the analog out is at the top of TB2
07:20 PM ithe2guu2795[m]: and how should i wire it to my 7i96s
07:20 PM djdelorie: it acts like a potentiometer, so follow the drive's directions for "external potentiometer"
07:20 PM ithe2guu2795[m]: like what pin out should i use from my dyn4 drive to the 7i96s?
07:21 PM ithe2guu2795[m]: i tried doing it before and was not able to get it wired up
07:24 PM djdelorie: Looks like you want interface type D, analog input on pin 13, ground pin 25
07:25 PM djdelorie: with the 7i96's spin+ going to +12v (+14v, maybe with some resistor to drop the max volts) and spin- to gnd
07:26 PM djdelorie: here's a thread that might help: https://forum.linuxcnc.org/pathpilot/47537-pathpilot-mesa-7i76e-dyn4-servo-drive-spindle-not-working
07:41 PM ithe2guu2795[m]: Ok, this is much harder to setup than i thought it would be
07:41 PM djdelorie: spindles can be a lot of wiring, yes
07:42 PM djdelorie: that's why I ended up going modbus
07:43 PM djdelorie: a lot of the mess is because they need to keep the various components electrically isolated
07:44 PM djdelorie: lots of opto isolators
07:44 PM * ithe2guu2795[m] uploaded an image: (317KiB) < https://libera.ems.host/_matrix/media/v3/download/jauriarts.org/WxfDLDimtGUcMOsBeGtXAMkW/image.png >
07:44 PM ithe2guu2795[m]: i am not able to translate this to a pinout for the 7i96s
07:44 PM ithe2guu2795[m]: some things just dont line up at all
07:45 PM djdelorie: spindle+ spindleout and spindle- are TB2 pins 22 to 24 on the 7i96
07:48 PM djdelorie: you could hard-wire the enable and direction inputs, or tie them to one of the 7i96's isolated outputs
07:48 PM djdelorie: you do not need to wire up the encoder lines just to use it as a spindle, if you're not doing threading
07:58 PM skunkworks[m]: djdelorie: mess? https://electronicsam.com/images/KandT/conversion/panel/electricalbox.jpg
07:59 PM djdelorie: no, mess: http://www.delorie.com/photos/cnc2/IMG_0649.html
08:00 PM skunkworks[m]: psha please.. you can count the number of wires in that picture.. 😉
08:01 PM djdelorie: yes but it's hard to trip over the wires in your box...
08:01 PM skunkworks[m]: lol - sure
08:05 PM -!- #linuxcnc mode set to +v by ChanServ
08:42 PM roycroft: i tidied up the wiring when i rebuilt my brew system control panel, and i think it turned out ok:
08:42 PM roycroft: https://roycroft.us/ControlPanelWiring/ControlPanelChassis.jpeg
08:42 PM roycroft: https://roycroft.us/ControlPanelWiring/ControlPanelFrontPanel.jpeg
08:43 PM roycroft: i really could have used that epson label printer with the shrink tubing labels when i did that
08:43 PM XXCoder: ;opls nice and tracable too
08:43 PM XXCoder: *looks
08:44 PM ZincBoy[CAON][m]: Very nice! Good use of connectors on the control panel.
08:44 PM roycroft: the original control panel did not have connectors
08:44 PM roycroft: it was a spaghetti wire mess going from the chassis to the panel
08:45 PM roycroft: and even though the connectors are all labeled, i also made sure to use a different type of connector for each cable, so i could not mix them up
08:46 PM roycroft: the mains going into that panel are 240v@50a, so i took care to do things as neatly and safely as i could
08:47 PM djdelorie: what did you end up using for an on/off switch?
08:47 PM roycroft: maybe i went a bit overboard with fusing things, but when you have high voltage/high current and lots of hot, highly conductive liquid in close proximity, it seemed prudent to isolate and fuse things as much as possible?
08:47 PM roycroft: me?
08:48 PM djdelorie: yeah. Do you have a 240v/50a switch, or using some contactor or otherwise?
08:48 PM roycroft: i used a momentary contact key switch to turn it on, and a mushroom e-stop switch to turn it off
08:48 PM roycroft: controlling a latching contactor circuit
08:49 PM djdelorie: ah. I was going to use contactors controlled by a lower-power circuit I could turn on/off normally
08:50 PM roycroft: so in that picture of the chassis, the third component from the left, with the red button at the top, is the latching contactor
08:50 PM roycroft: right next to that on the right is a 63a power distribution block
08:50 PM roycroft: then to the right are circuit breakers for the individual circuits that it controls
08:51 PM roycroft: to the right of the breakers is a 12vdc power supply
08:51 PM roycroft: the pid controllers managed a couple high current ssrs mounted on the back panel
08:52 PM roycroft: but since ssrs tend to fail closed, i have some contactors in the main chassis as well
08:52 PM roycroft: so when i engage a heating element, the main contactor supplies current to the ssr
08:52 PM djdelorie: for my cnc I'm trying to figure out if I need to worry about cycling the various power supplies in the right order...
08:53 PM djdelorie: or just CLUNG turn them all on at once
08:54 PM roycroft: i don't seem to have any pictures of the rebuilt control panel online, but this is the original panel, which is very similar:
08:54 PM roycroft: https://roycroft.us/BrewSystem/IMG_1012.JPG
08:54 PM djdelorie: very retro-pretty :-)
08:54 PM roycroft: yeah, i was going for the retro look
08:54 PM djdelorie: needs a dummy switch labelled "Launch"
08:55 PM roycroft: the wood is mahogany, and i spent a lot of time getting the color and finish just right
09:06 PM XXCoder: excellent
09:52 PM mrec: rmu: the question would be which setting :-) I sent all the information I have to the endor JMC directly let's see if we get something back
11:22 PM -!- #linuxcnc mode set to +v by ChanServ